The primary goal of this protocol is to characterize adrenal function and to investigate the dose-dependent effects of glucocorticoids during vasopressor dependent septic shock. Septic shock is responsible for approximately 400,000 intensive care unit admissions and 200,000 deaths each year in the United States. (1) The identification and development of therapies that can improve survival in patients with septic shock is essential. Glucocorticoids (steroids) have been investigated as a therapy for sepsis over the last 40 years with variable results. Glucocorticoids act on nearly all cells in the body, with effects that are especially important to the stress response. They increase the synthesis and action of catecholamines which increase cardiovascular contractility and blood pressure. Glucocorticoids also modulate the immune and inflammatory responses. Initial clinical trials using high dose glucocorticoids to suppress an excessive inflammatory response demonstrated that steroids may have adverse effects on survival. These harmful effects may have been due to the increased risk for secondary infections with the immunosuppressive effects of steroid therapy. However, these early trials enrolled a patient population with varying levels of severity of illness from mild sepsis i.e. presence of infection with fever, tachycardia, and normal blood pressure, to severe septic shock, i.e. sepsis with hypotension and organ hypoperfusion requiring vasopressors to maintain blood pressure referred to here forward as "vasopressin-dependent septic shock". (2) This range in severity of illness may also, in part, account for the harmful effects of glucocorticoids seen in some patients in the early clinical trials. We have developed a sedated and ventilated model of canine Staphylococcus aureus bacterial pneumonia that simulates many of the pathophysiologic changes occurring during clinical sepsis. Most importantly, despite volume resuscitation, these animals develop a persistent vasopressor requirement within hours of the onset of bacterial pneumonia. An intra-bronchiole bacterial load of 3.6-4.0 x 109 cfu's/kg has been found to result in about a 50-60% mortality. This mortality range will allow us to determine if treatment with steroids is beneficial during sepsis. This clinically representative model of sepsis will allow us to better understand the physiology of adrenal function in septic shock and to determine 1) how adrenal function changes over time during sepsis, independent of steroid therapy; 2) the relationship between adrenal function and total and free cortisol levels as well as more sophisticated tests of the hypothalamic-pituitary-adrenal axis dysfunction and outcome; and 3) the mechanism of steroids? effects at low and high doses on survival and shock reversal during vasopressor-dependent septic shock.